requirement-quality
Pass
Audited by Gen Agent Trust Hub on May 26, 2026
Risk Level: SAFENO_CODEPROMPT_INJECTION
Full Analysis
- [SAFE]: No malicious patterns or security risks were identified. The skill consists entirely of instructional Markdown content and reference definitions.
- [NO_CODE]: The skill contains no executable code, scripts, or binaries. It operates purely as a set of logical guidelines for the agent to follow during text processing.
- [PROMPT_INJECTION]: The skill exhibits a surface for indirect prompt injection as it ingests untrusted feature specifications. 1. Ingestion points: Feature specifications provided by the user and configuration files (.lattice/config.yaml). 2. Boundary markers: None. 3. Capability inventory: No tool use or executable scripts across provided files. 4. Sanitization: None. Due to the lack of dangerous capabilities, the risk is negligible.
Audit Metadata